Cette contribution constitue la première approche systématique des modalités d’expression de la référence impersonnelle humaine en langue des signes française (LSF). Elle prolonge une précédente investigation que nous avions menée sur la base de l’analyse de corpus de discours (L’Huillier et al., 2016, Garcia et al., à paraître). La description s’appuie d’abord sur les données vidéo élicitées obtenues à partir d’un questionnaire sur la référence impersonnelle humaine initialement élaboré pour des langues vocales (LV), puis adapté pour les langues des signes (LS) par Barberà & Cabredo Hofherr (à paraître) dans le cadre d’un Projet franco-allemand (Towards a typology of human impersonal pronouns ou TypoImp). Les stratégies identifiées sont ensuite mises en regard de celles que nous avions dégagées dans nos corpus. Après une brève revue de la littérature sur la question et une présentation de notre cadre théorique, nous explicitons la méthodologie adoptée et les questions soulevées par la mise en place du protocole d’élicitation, de la préparation du questionnaire à la représentation des données. Nous exposons finalement, puis discutons, les principales stratégies d’expression de la référence impersonnelle humaine que nous avons mises en évidence pour la LSF.
The study presented in this paper is a first systematic approach to the expression of the impersonal human reference in French sign language (LSF). It extends and deepens a prior study carried out by the authors on the basis of a large scale discourse corpus. The description proposed here is based primarily on data elicited through a specialised questionnaire on impersonal human reference (presented in Barberà & Cabredo-Hofherr to appear), initially developed for spoken languages (SpLs) and adapted for sign languages (SLs). The strategies revealed are compared with those discussed in our prior study. We begin with a brief review of the literature on impersonal human reference in SpLs and SLs, and a presentation of our theoretical framework for the analysis of LSF. We then elaborate on our methodology and the issues raised by the elicitation protocol adopted, from initial stages of its preparation to the representation of our data. As for our main results, our data exhibit the structuring character of the existential/universal (unrestricted) opposition in LSF ; the first tend to occur with overt agent marking, while the preferred strategy in the latter appears to be null subject. There is also a clear contrast in existential contexts, between a singular agent (marked by ONE (SOMEONE) and those that evoke a non-singular agent or a vague plurality (marked by what we glossed Circ Pt / 3pl-On). However, the existential vs. universal opposition must be nuanced. Existentials with a generalizing / habitual predication clearly involve the idiomatic sign PI, which is a strong habitual/typicality marker in LSF. A central feature of this marker seems to be the exclusion of an overt impersonal agentive subject. This case is illustrated in locative universals, which also indicate typicality (an attitude or characteristic of a particular group), and exhibit very similar strategies to signal impersonal human reference (null subject associated with PI). The null subject strategy, the preferred strategy in unrestricted universals, seems to be a possible alternative in most contexts, existential or universal. Setting aside conditionals and verbs of saying, the one exceptional context in this respect, notable but semantically consistent, is the anchored (episodic) existential which seems to obligatorily require overt marking of an impersonal agent. Our elicited data confirmed an observation made in our previous discourse corpus study, namely : the highest degree of impersonality (in unrestricted universals) is expressed in LSF through subject ellipsis without any spatial anchoring for the lexical sequence of the utterance (neutral space). In addition, we have identified two particularly interesting impersonal markers used in universal contexts : the 2nd person pronoun and the so-called “instructional personal transfer” (“TP prescriptif”). The 2nd person pronoun as a marker of impersonal reference (IX2), although rare in our data, has been identified in other SLs, including LSC (Barberà & Quer, 2013). The instructional personal transfer, also rare in our elicited data, is prevalent in our discourse corpora, and appears to be restricted to specific semantic contexts—the presentation of a sequence of instructions, a manufacturing process or a recipe. We have suggested to relate this specific type of personal transfer to the impersonal 2nd person used in some SpLs, particularly in their oral (non-written) form. In contrast, we note that our data did not include a number of strategies observed in other SLs, such as the impersonal 1st person pronoun (IX1) or the interrogative pronoun WHO. The latter is attested only once in our elicited data (in the vague existential context), but was preceded by the signs ONE (SOMEONE) and Circ Pt / 3pl-On, in contrast to its use in LSC, for example. Another significant point that must be addressed is the use of the upper area of the frontal plane as a marker of impersonal human reference in LSF. We have noted that this location has been identified as a central impersonal marker in LSC, which echoes past observations with respect to LSF. It should be noted, first, that the relevance of this high ipselaterale zone was noted in our discourse corpus study, where we identified it as the starting point of agreement verbs such as SAY, INFORM or GIVE used to indicate an undefined/impersonal human agent (commonly translated by French on). Although not discussed in this article, we must stress the importance of the 'high zone' marker in LSF for constructions of this type (indefinite agent with agreement verbs). Moreover, in our elicited data, the high zone was attested in only two types of contexts, the anchored (episodic) existential and the existential with plurality of referents (vague plurality). In the first case, which also excludes the use of the null subject, our data lead us to conclude that the anchoring of a pronominal or nominal agent ONE (SOMEONE) or PERSON in the upper area of the frontal plane is not related to impersonal marking. Rather, it is motivated by iconic considerations, as this location topographically reflects the high location of the agent relative to the object operated (the bell /the light flash). In contrast, the upper area associated to the Circ-Pt / 3pl-On marker in existentials with vague plurality clearly conveys the low referentiality of the agent. This is one of the points that await further confirmation via systematic study of our discourse corpora. Finally, some comments may be added with respect to the null subject strategy. At this stage, we can offer two hypotheses, which also require further research. First, we have observed that the null subject strategy is an available option in almost all contexts tested by the questionnaire. We noted that, in the vast majority of cases, the target sentences could be translated to French with the pronoun on. This pattern corresponds to analyses of French on, namely that it covers all impersonal uses identified so far in SpLs (see Gast & van der Auwera, 2013). As such, it seems reasonable to hypothesise that the null subject in LSF is the closest equivalent to French on, and not the circular pointing forms (annotated Circ-Pt / 3pl-On for lack of a better gloss), although the latter would appear close to French on in some of its values. In our analysis of the Circ-Pt / 3pl-On marker, we have raised the possibility, left for future research, that it may be equivalent to the 3pl-IMP markers (i.e. impersonal use of the 3rd person plural personal pronoun) described in the literature on SpLs, or that it is a marker unique to LSF. The second hypothesis directly concerns a typological issue. In the absence of systematic study of this issue, we cannot say for certain whether LSF should be considered a pro-drop (or partially pro-drop) language. However, our empirical observations suggest this is indeed the case. If so, our results in this study corroborate the Siewierska’s (2011) correlation between pro-drop languages and the null subject strategy.
